Morfessor EM+Prune: Improved subword segmentation with expectation maximization and pruning
Autor: | Grönroos, Stig-Arne, Virpioja, Sami, Kurimo, Mikko |
---|---|
Přispěvatelé: | Calzolari, Nicoletta, Bechet, Frederic, Blache, Philippe, Choukri, Khalid, Cieri, Christopher, Declerck, Thierry, Goggi, Sara, Isahara, Hitoshi, Maegaard, Bente, Mariani, Joseph, Mazo, Helene, Moreno, Asuncion, Odijk, Jan, Piperidis, Stelios, Dept Signal Process and Acoust, University of Helsinki, Aalto-yliopisto, Aalto University, Calzolari [et al.], Nicoletta, Language Technology |
Jazyk: | angličtina |
Rok vydání: | 2020 |
Předmět: |
FOS: Computer and information sciences
Morphology Tools Computer Science - Computation and Language Language Modelling 6121 Languages Statistical and Machine Learning Methods 113 Computer and information sciences Computation and Language (cs.CL) Less-Resourced/Endangered Languages Unsupervised learning |
Popis: | Data-driven segmentation of words into subword units has been used in various natural language processing applications such as automatic speech recognition and statistical machine translation for almost 20 years. Recently it has became more widely adopted, as models based on deep neural networks often benefit from subword units even for morphologically simpler languages. In this paper, we discuss and compare training algorithms for a unigram subword model, based on the Expectation Maximization algorithm and lexicon pruning. Using English, Finnish, North Sami, and Turkish data sets, we show that this approach is able to find better solutions to the optimization problem defined by the Morfessor Baseline model than its original recursive training algorithm. The improved optimization also leads to higher morphological segmentation accuracy when compared to a linguistic gold standard. We publish implementations of the new algorithms in the widely-used Morfessor software package. Accepted for publication in LREC 2020 |
Databáze: | OpenAIRE |
Externí odkaz: |